Single Tube PAS* Linestage
Posted: Thu Apr 14, 2011 1:37 am* = or any other linestage 
Hi Folks,
Haven't posted a recent project, so I'll post another commission here.
Goal: a single tube linestage for the PAS2/3/3x preamplifier. Cannot infringe on I.P. or topology of the Curcio single tube pre.
Challenge: Getting gain control and high performance out of a single tube.
Solution: Hybrid anode-mu-follower.
With an anode follower, the gain will float about on varying driving and load impedances. So, that had to be isolated.
I'm still on a bit of a hybrid-mu kick (especially since the PAS power supply is about as strong as a housefly), and the source follower operates very well at low currents, like a 12AX7/5751/6SL7 like, so that was the top stage as a natural choice.
But rather than taking the feedback from the output, I contained it before the output follower. I found an advantage for this for a few reasons:
- Better isolation with heavy loads (which would make the gain go up), as found in many cheaper SS amps.
- Less chance of becoming an oscillator with non-resistive lods.
- Noise from the load wasn't coupled back into the grid circuit.
Anode followers are kind of like an op-amp, in that the gain varies from the ration of the grid resistor to anode FB resistor. Better performance and sonics is had from a lower grid resistor, but at the expence of becoming increasingly difficult to drive.
I settled on 100K for the grid resitor. Not too hard of a load to drive, but when using a volume control up front, strange things happened and the gain wasn't linear, nor the frequency response. So, we needed to isolate that too. A source follower seemed a logical choise.... NOT!
The source follower was phenomenally noisy. STP2NK60Z was noisy, but not near as bad as an IRF-anything
After chatting with a colleague and getting tips on reducing noise, I was able to lower it, but not eliminate it. For the heck of it, I tossed in a bipolar. The noise was gone!
That led to the development of this baby:
A neat thing about the fixed-bias nature of this, is you can swap in a 12AX7 instead and not have the DC voltages shift all over the place (it doesn't sound too bad, either). With the anode follower fixing the gain, the higher gain tube doesn't affect the overall gain of the circuit. F3 changes a bit due to Cmiller, but is mostly negated by the large, unbypassed cathode resistor.
As long as the B+ stays in a range where the zener diodes work, gain, waveform performance and to my ear, sonics, aren't affected by even large B+ swings. Positive side effect of the mu-follower arrangement.
Another side effect that I see as a positive (many might see as a negative), is the topology makes it rather immune to tube rolling. Your RCA or Philips cheapie will sound the same as a Telefunken or Mullard in this.
Forgot to mention about Ccomp - it was chosen for best square wave... maybe a little sharp even on the leading edge (as in a nipple, not railroad spike). This too, once chosen, rolling tubes has little effect on the waveform.
Cheers!
Hi Folks,
Haven't posted a recent project, so I'll post another commission here.
Goal: a single tube linestage for the PAS2/3/3x preamplifier. Cannot infringe on I.P. or topology of the Curcio single tube pre.
Challenge: Getting gain control and high performance out of a single tube.
Solution: Hybrid anode-mu-follower.
With an anode follower, the gain will float about on varying driving and load impedances. So, that had to be isolated.
I'm still on a bit of a hybrid-mu kick (especially since the PAS power supply is about as strong as a housefly), and the source follower operates very well at low currents, like a 12AX7/5751/6SL7 like, so that was the top stage as a natural choice.
But rather than taking the feedback from the output, I contained it before the output follower. I found an advantage for this for a few reasons:
- Better isolation with heavy loads (which would make the gain go up), as found in many cheaper SS amps.
- Less chance of becoming an oscillator with non-resistive lods.
- Noise from the load wasn't coupled back into the grid circuit.
Anode followers are kind of like an op-amp, in that the gain varies from the ration of the grid resistor to anode FB resistor. Better performance and sonics is had from a lower grid resistor, but at the expence of becoming increasingly difficult to drive.
I settled on 100K for the grid resitor. Not too hard of a load to drive, but when using a volume control up front, strange things happened and the gain wasn't linear, nor the frequency response. So, we needed to isolate that too. A source follower seemed a logical choise.... NOT!
The source follower was phenomenally noisy. STP2NK60Z was noisy, but not near as bad as an IRF-anything
After chatting with a colleague and getting tips on reducing noise, I was able to lower it, but not eliminate it. For the heck of it, I tossed in a bipolar. The noise was gone!
That led to the development of this baby:
A neat thing about the fixed-bias nature of this, is you can swap in a 12AX7 instead and not have the DC voltages shift all over the place (it doesn't sound too bad, either). With the anode follower fixing the gain, the higher gain tube doesn't affect the overall gain of the circuit. F3 changes a bit due to Cmiller, but is mostly negated by the large, unbypassed cathode resistor.
As long as the B+ stays in a range where the zener diodes work, gain, waveform performance and to my ear, sonics, aren't affected by even large B+ swings. Positive side effect of the mu-follower arrangement.
Another side effect that I see as a positive (many might see as a negative), is the topology makes it rather immune to tube rolling. Your RCA or Philips cheapie will sound the same as a Telefunken or Mullard in this.
Forgot to mention about Ccomp - it was chosen for best square wave... maybe a little sharp even on the leading edge (as in a nipple, not railroad spike). This too, once chosen, rolling tubes has little effect on the waveform.
Cheers!
